relax SVN archive

The multi-processor data_fetch() API function is now used to obtain the total_length variable.

Fix for the Processor.data_upload() method.

The self.run_command_queue() method is no longer being used. Instead the add_to_queue() method is
being used to queue up the slave commands, and the run_queue() method is being called at the end to
flush the queue.

Fixes and clean up of the slave to master transfer mechanism for result commands.

The master_queue_command() and master_receive_result() methods are now defined at the Processor
level (but call raise_unimplemented()). Full epydoc docstrings and comments have been added
everywhere. The 2 methods have now been defined for uni-processor fabric, and simply hold the
command in a temporary variable.

Shifted the self.threaded_result_processing flag into the base Processor class where it belongs.

Clean up and completion of the TODO for the Processor.assert_on_master() method.

The Processor.assert_on_master() method has been created and calls raise_unimplemented(). The
Multi_processor.assert_on_master() method has been shifted to Mpi4py_processor.assert_on_master(),
as that method's error message is MPI specific. The empty Uni_processor.assert_on_master() method
has been added to allow that fabric to work.

Whitespace fix.

Spun out all of the results queue objects into their own module.

This completes another set of TODOs by removing these queue objects from any fabric level. They
can now be imported and used by any fabric level (Processor, Multi_processor, Mpi4py_processor,
Uni_processor, etc.).

Shifted the run_command_queue() and run_queue() methods from the Multi_processor to Processor class.

The multi/test_implementation2.py script can now be run in uni-processor mode.

Thought it still fails due to deficiencies in the Processor base class (one of Gary's TODOS in the
Multi_processor class).

Comment fixes.